AI curriculum
Course outline for Retrieval-Augmented Generation
Generated answers can omit or misrepresent relevant evidence. Build a retrieval pipeline and test whether its context supports answers to a defined set of questions.
About Retrieval-Augmented Generation
Generated answers can omit or misrepresent relevant evidence. Build a retrieval pipeline and test whether its context supports answers to a defined set of questions.
Retrieval-Augmented Generation Course Objectives
- Build a document chunking and retrieval pipeline.
- Compare vector, hybrid, and reranked results.
- Diagnose grounding failures with retrieval tests.
Pre-requisites
- Basic programming and data handling skills.
- Familiarity with LLM application inputs and outputs.
Lab Setup
- Computer with a programming runtime and a local retrieval library.
- Use public or synthetic documents and local embeddings or supplied vectors.
- Local model or response fixtures; no hosted vector service is required.
Detailed Course Outline
Proposed modulesModule 1: Preparing evidence
- Why grounding is needed
- Chunking
- Embeddings
- Vector search
Practical outcome: Retrieve relevant chunks from a small document set.
Module 2: Constructing context
- Vector indexes
- Hybrid search
- Reranking
- Context construction
Practical outcome: Compare retrieval strategies using the same queries.
Module 3: Testing grounding
- Retrieval evaluation
- Graph RAG
- RAG failure modes
Practical outcome: Identify missing evidence and unsupported answers.
Practical exercise
- Build a grounded question-answering prototype and document retrieval failures on a test set.
How we train
Contact us for full course details, including duration, delivery options and lab requirements.
Ask about exercises, instructor feedback and the prior knowledge you need. Tool-specific courses marked provisional may become modules in a broader course.
Discuss your learning goals